Assessing the PR of a backlink

Hello Everyone,

This is my first post, so hopefully it's not too redundant.

Often when choosing which sites to create backlinks on first, I look at the PR to try and get a feel for the amount of juice that will flow to my website. The question is, do I measure the PR of the domain or the PR from the individual web page? Many domain names carry a PR of 4 or 5, but contain many different sub-category web pages, from which I am linked. Often this specific, sub web page does not show up on a PR ranker.

Best,
Mike

I would say the page rank of the page you are linking from is more indicative and immediately relevant, but the PR of the sites root/home page (or of the page right above the linking page in the site's internal linking hierarchy) also plays a considerable role. Most importanly, a link from even a minor page of a highly trusted site is likely to be excellent for your site's TrustRank.

Remember that PageRank is *not* the sole decisive factor in link value. There is also TrustRank, equally if not more important. Somehow Google's PageRank and TrustRank algos work together.

A link from a .edu or a .gov site with no to little page rank may do your rankings more good that a link from a dot-com with a high PR.

IN ADDITION to the linking page's "toolbar PR", consider these factors that come to mind:

(1) The site's overall quality and relevance to your field or market. If it's an excellent and relevant site, a link from to will do you good regardless of its page rank. (Additionally, good sites' PR tends to increase with time, so "link equity matures.")

(2) Is the page cached? If the page is not indexed, forget it. Also, check the date of the last cache. If the cache is older than 45 days, they say the page doesn't command much search engine trust or carry much SEO value.

(3) The number of inbound links on the page you are linking from (the more quality inbound links the better) and the number and quality of outbound links on the same page. (Among the alleged purposes of the TrustRank algo is the function of flagging sites with high PageRank that have no *trusted* link equity, their PR being presumably accumulated through spam. Therefore a link from a site with a high toolbar PR is *not always* what you want.)

(4) Avoid sites that manifestly sell links (look for such clues as the mention of "paid listings," "sponsored links," "advertise," etc.). These no longer seem able to pass any Google PR.

(5) If the link contains rel="nofollow" it will have zero SEO value to you, thought it may still drive traffic.

(4) If you need to get a link on a site that has no trust, *make sure* it is nofollowed, so it doesn't hurt your SEO (this reassures the search engines that you did not buy for the link). But that's a special case.

(5) Make sure that the link you are getting is a good text link where you can control the anchor text (or the Alt tag if you must use an image), and that the href attribute contains a straightforward simple URL rather that being implemented with a redirect. (Some redirects are actually fine and can be recognized by the Search Engines as links to your site, but as a general proposition they are best avoided in favor of simple links.)

(6) If you can, ensure that the page you are linking from includes some optimization (such as perhaps including your company/product name in the URL, title tag, <h#> tags etc.).

So those are the six "cents" I felt inspired to share. I hope they are of some interest.

Other SEOs, please feel free to add to this list!

Trustrank has nothing to do with Google's algorithm.

A link from an edu or a gov site is of no greater importance than a link from a .com, all other things being equal of course.

The rest is pretty solid and nicely written imo.

If it's a site where you have the ability to create your own links, then most likely you're not spending your time all that wisely.

PR of the linking page is more important than the PR of the linking domain. This is because ur site will get the PR of linking page and not that of the root domain.

Even if the linking page has no PR it can still be targeted provided it is frequently crawled and indexed. Every time a page is crawled its PR is calculated and then it comes up on the PR toolbar by next update. But in that case root domain should have good PR and the linking page should not be several directories deep.

Thank you for your kind observations.

About TrustRank: not so. The name seems to be used in two different (though related) ways. First of all, TrustRank is the method proposed by some scientists at Stanford to detect search engine spam. More recently, TrustRank is a trademark and a patent owned by Google. Google appears to have a TrustRank algo. You can find out more here
http://www.seobythesea.com/?p=3015 and elsewhere.

Regarding links from .gov and .edu sites not being any different from .coms -- these thing are speculative and based SEOs' experiences. Like a lot of SEO, it's pure observation. I am pretty sure you have no hard evidence and can't site a statement by Google that all these links are equal in status. So why such confidence?

You can have a crap site on a .edu just as well as you can have a crap site on a .com.

It is not the tld that makes the difference, but rather the quality of the site hosted there that matters.

To say that a site is trustworthy solely because it is on a .edu would be negligible (blanket statement). A better statement would be to say to get links from educational institutions (which can also be on .com .net or whatever they chose at the time of registration).

TrustRank is an invention by Yahoo brother. Google trademarked it to.... well I'll just let Cutts fill you in.

Trustrank explained

Further:



http://www.stephanspencer.com/search-engines/matt-cutts-interview

Now, I am one to say "trust, but verify" especially when it comes to Google. So I did. Please don't make me find my research from 2 -3 years ago




"Pretty sure" is doing you no favors dude.

This is a statement nicely based on common sense. I don't disagree with the sense behind it, but how things actually work is an empirical matter completely independent of common sense. For what it's worth, many seem to report that links from any .edu pages, even completely obscure and unimportant ones, such as the personal pages of students, employees or alumni, are excellent for your SEO. This *may* be a myth, but I haven't seen any evidence yet to suggest that it is. Your mileage may vary.

Have you seen any evidence yet to suggest that it is true?

Just because you can't prove something is a myth doesn't make it reality.

KernelPanic, OK I see, but SEO-related statements and evidence from 2-3 years ago strike me as pretty old and no longer reliable when it comes to Google, which is always a moving target. A lot of water has flowed under the bridge. Example: nofollow as been to hell and back in the past couple of years (or has it been more?). They says the algos change every week.

Well, I have no conclusive evidence, but I have heard/seen this statement made by seemingly reasonable people that I don't mistrust for independent reasons.

Come to think of it, a similar uncertainty haunts, I would say, over 30% of the SEO that I know, but I doubt that all of that is a myth (and sure do hope that it's not).

On a more commonsensical note, .edu and .gov sites are typically non-commercial or not overtly commercial, and such domains cannot be purchased and *are* known to generate relatively little spam compared to the .com. It does not seem entirely absurd for Google to trust such domains more, if that's indeed what it does.

Typically, yes. Always, no.

And therefore, it's not the tld that matters to Google, but the site hosted there.

I agree that that's what matters. However, Google's evaluation of site quality is based on formalizable criteria, and I would say ease of computation plays a role. The .edu and .gov sites have already been vouched for by an accredited university or the US government, and those are pretty good vouchers, so you might say they start out with some advance credibility.

That said, I have every intention to try getting .edu links to this or that project of mine at some point in the future, just to see the effect, if any. Perhaps there are people here who have already tried and seen? Do share!

So are you trying to tell me that Google has decided to use Yahoo's Trustrank factor in their algorithm afterall?

You should let this one go.